Data Integrity
Data integrity refers to the completeness, consistency, and accuracy of data throughout its entire life cycle. In metrology, if the integrity of calibration data is compromised, the traceability chain is broken, and all downstream measurements relying on that calibration become suspect.
The ALCOA+ Principles
The standard framework for ensuring data integrity across regulated industries (especially pharmaceuticals and aerospace) is known by the acronym ALCOA+. A measurement record must satisfy all of these conditions to be considered trustworthy.
Core ALCOA Attributes
- Attributable: It must be completely clear who observed and recorded the data, and when it was done. Every data point or modification must link back to a uniquely identified individual (e.g., via secure electronic signatures, not shared login credentials).
- Legible: The data must be readable, understandable, and accessible throughout its required retention period. This includes ensuring that proprietary file formats can still be opened years later.
- Contemporaneous: Data must be recorded at the exact moment the observation is made. Relying on memory and writing down results later is a severe violation of data integrity.
- Original: The first capture of the data (whether on a paper datasheet, printed from a thermal printer, or saved to a database) is the original record. Exact copies can be made, but the original must be preserved. Transcribing data introduces risk.
- Accurate: The data must be correct, truthful, valid, and reliable. This ties directly back to equipment calibration, proper metrological procedure, and environmental controls.
The '+' in ALCOA+
As digital systems evolved, regulatory bodies added four additional attributes to clarify expectations for electronic records:
- Complete: All data, including repeat tests, failed calibrations, and metadata (like temperature during the test) must be retained. "Cherry-picking" passing results while discarding failing runs is data falsification.
- Consistent: Data must be structured and chronological. The sequence of events should make sense (e.g., an "As-Left" measurement cannot logically occur before an adjustment).
- Enduring: Data must be stored on reliable media designed to last for the duration of the retention period (e.g., using enterprise-grade servers rather than individual flash drives).
- Available: Data must be readily available for review, audit, or inspection by regulatory bodies over the lifetime of the record.
Audit Trails
A cornerstone of electronic data integrity is the audit trail. This is a secure, computer-generated, time-stamped electronic record that independently tracks the date and time of operator entries and actions that create, modify, or delete electronic records.
If a technician corrects a typo in a measurement result within a LIMS, the audit trail must record the original value, the new value, the identity of the person making the change, the exact timestamp, and the reason for the change. The audit trail itself must be secure from tampering, even by system administrators.